try the following link:

http://protegewiki.stanford.edu/wiki/Ontology_Evaluation

What is the metrics you want your ontology to be evaluated on?

Hello Asim. I suggests that you can use one (or both) of this options:

1. Ontology Pitfall Scanner (OOPS!) http://oops.linkeddata.es/advanced.jsp

2. OntoUML Lightweight Editor. https://code.google.com/p/ontouml-lightweight-editor/

I think this options are really good. I hope you find this usefull.

I have also tried OOPS! suggested by Juan and found it very good!

Chris is correct, you need to be more specific on the kind of evaluation you are looking for. For instance, to check the consistency of your ontology, you might use a reasoner, assuming your ontology is in a decidable language like OWL. If you are more interested in evaluating the coherence of your representation according to domain knowledge, then you might need to work closely with domain experts.

In order to compute ontology metrics in general I suggest you use my Java API called OntoMetriCalc. OntoMetriCalc is a Java API for measuring OWL ontologies based on Jena. It is easy to use and supports 29 ontology metrics including: Maximum Depth of Class Inheritance Tree, Maximum Number of Class Ancestors, Average Depth of Class Inheritance Tree and Average Number of Class Ancestors.

You can find OntoMetriCalc here:

http://sourceforge.net/projects/ontometricalc/

Just write this code in Java using OntoMetriCalc to measure an ontology:

public void measureOntology() {

MetricsCalculator calc = new MetricsCalculator("ontology.owl");

Iterator metrics = calc.getAllOntologyMetrics().iterator();

while (metrics.hasNext())

{

Metric metric = metrics.next();

if (metric.getType()==0)

System.out.println(metric.getName()+": "+metric.getIntValue());

else

System.out.printf("%s: %.2f\n", metric.getName(),metric.getDoubleValue());

}

}

The precision and recall metrics can't be computed by OntoMetriCalc. The reason for this is that precision and recall are two metrics from information retrieval. It is true that you can also measure ontologies, in particular ontology alignments, using precision and recall. However these metrics can't be computed just like that. Information regarding what is considered to be relevant or irrelevant must also be present. This is a matter of individual perception and thus it is very subjective. However, in logical terms, precison essentially means correctness and recall essentially means completeness.

Thus, one thing you could do is to evaluate your ontology for these two criteria (correctness and completeness) and you can use OOPS! for this. OOPS! has a criterion called Completeness (in the web interface go to advanced evaluation). The Correctness criterion is not explicit. Because of this I think it would be even better if you use DrOntoAPI (another Java API that I created and this one uses the OOPS! web service) which has both these criteria (and others) or you can even use it to define your own criteria and place the OOPS! pitfalls that you think are specific to these criteria there. This allows you to receive an evaluation from OOPS! that is specific to your criteria.

Just write this code in Java that uses DrOntoAPI to evaluate the GoodRelations ontology with OOPS! for correctness and completeness:

public void evaluate() {

//Create a client that communicates with the OOPS! web service provider.

IWebService webService = new WebService();

try {

//Set the ontology that will be evaluated.

webService.setOntologyIRI("http://purl.org/goodrelations/v1");

//webService.setOntologyFile("v1.owl");

//Select only some pitfalls

PitfallSelector pitfallSelector = new PitfallSelector();

//Select pitfalls based on a filter rather than directly

IFilter filter = new Filter();

filter.filterPitfalls(Criteria.CORRECTNESS);//filter pitfalls by correctness

filter.filterPitfalls(Criteria.COMPLETENESS);//also filter by completeness

pitfallSelector.selectPitfalls(filter);//add filtered pitfalls to the selection

//tell the web service which pitfalls to check based on a PitfallSelector

webService.setPitfallSelector(pitfallSelector);

//View selected pitfalls

System.out.println(pitfallSelector.toString());

//Create a model of the response to manage it

IPitfallManager manager = new PitfallManager(webService.getResponse());

//Generate a CSV report of the evaluation result

manager.generateReport(new FileOutputStream(new File("csv_report.csv")), Report.CSV_REPORT);

}

catch (IOException | UnexpectedErrorException e)

{

e.printStackTrace();

}

}

You can find DrOntoAPI here:

http://sourceforge.net/projects/drontoapi/

For more clarifications, you can use the Forum from DrOntoAPI's SourceForge page.

In other words, what I am suggesting you is to build your own tool using DrOntoAPI or OntoMetriCalc or both. It is easier than it seems and you have total freedom rather than being limited to what a tool has to offer.

Good luck!

Thank you the respected researchers for the valuable support. Thank you all. 

Anyone pls tell me about Ontology Evaluation Tool. Onto-metric is not working online. Pls help me.

Sonika Malik

This online evaluation tool works fine:

https://ontometrics.informatik.uni-rostock.de/ontologymetrics/

It appears that both OntoMetrics and OOPS! are both down as of today (both give 503/502 errors). I can't contact the OntoMetrics creator as the contact tab isn't working either (I am contacting the OOPS! email and will update if they respond). Does anyone know if there are plans to bring these tools back online?

EDIT 1: Contacted OOPS! and working through options. I built my ontology (the GSSO) in Protege, so I was hoping that the Ontology Evaluation package (https://protegewiki.stanford.edu/wiki/Ontology_Evaluation, listed above) might be an option, but it hasn't been updated since 2013 and is only compatible with Protege 3.4.8 and 4.1 (I'm currently running Protege 5.5.0).

EDIT 2: OntoMetrics works after I manually removed all instances from the ontology!

EDIT 3: OntoMetrics was able to run everything behind-the-scenes (the ontologies I was running were a bit too big). OOPS! also got back to me and quickly fixed an encoding issue with OnToology [http://ontoology.linkeddata.es/] which now works fine!